CLAIRO confirms that sensor networks open new fields in air quality monitoring

Sensor technology gaining ground in air quality monitoring

As opposed to very expensive and large traditional regulatory instruments the price of which can be at least USD 10,000, low-cost air quality sensors, which can be purchased for the fraction of that price, offer a wealth of opportunities for local authorities when the goal is to undertake accurate, high-resolution measurements. These small devices, which are often portable or wearable due to their reduced price, have become recently widely available for the general public. Low-cost air sensors have multiple advantages over conventional air quality monitoring stations. They allow simple installation and use and retain proper functionality over long time with little or no maintenance. Their easy use enables the monitoring of remote or inaccessible areas. Thanks to their availability, they can be deployed in a large number and connected into clusters to monitor closely an area of interest. They allow fast measurements in dense networks. Sensors can also be linked to official national monitoring networks, adding new dimensions to data collection. Thanks to miniaturization, sensor units can be as small as a mobile phone. Finally, measurements by the sensors do not need to be performed by trained personnel and the monitoring process does not involve laboratory analysis.

Despite the fact that sensor technology has undergone great development in recent years, there are still crucial technical limitations linked to sensory measurements. As compared to traditional regulatory instruments air quality sensors have lower measurement precision, accuracy, and long-term signal stability. Due to the degradation of sensors, regular calibration of the measured values is necessary to improve the accuracy of the devices. When sensors are calibrated with the help of conventional air quality monitoring stations, correction factors are developed that are used to validate the data measured by the sensors.

The sensor technology has emerged 25 years ago and during the last 10 years its use became more and more common. The first generation of sensors was used as detectors to indicate high concentration of carbon monoxide in mines and later in houses and flats to prevent poisoning. Another early application of the technology was to monitor the concentration of methane and oxygen in mines to indicate explosiveness and flammability. As a result of the miniaturization efforts, the reliability and accuracy of sensor measurements remained relatively low. At the same time, over time with the price falling, mass production became possible, and as a result sensors could be purchased and used by the general public. It quickly became evident that it would make sense to connect the sensors in networks to monitor selected areas or zones more closely. Owing to their small size and low energy consumption units can also be used as portable devices.

The use of sensors today is manifold. They are used as professional detectors at production facilities, as components of dense networks of sensing units to allow the development of datasets, maps or visualizations. In some applications units are mounted on vehicles of public transport, and these circulating units can cover a wide area, performing alone the work of many stationary units. The technology also supports community-based, or citizen-science based initiatives, under which volunteering individuals are required to assemble and use the devices.

Dense sensor networks deployed in Ostrava

The City of Ostrava has embraced the sensor technology in the CLAIRO project as it offers high resolution data sourced from relatively small areas, and because of the large number of monitoring points, cost was a crucial factor. Access to detailed data was necessary for the partnership to be able to assess impact of urban greenery on pollutant removal and to make predictions for future capture.

Altogether 19 sensor units and one reference system were deployed by the Technical University of Ostrava (VSB) in the two target areas of the CLAIRO project. The Radvanice district in Ostrava, where the target sites are located is one of the most polluted areas of Czech Republic. Intentionally, sites with poor air quality were selected to be able to demonstrate significant air quality improvements with the use of greenery.

Sensor boxes that contain several sensors are hung on 4-meter-high poles. They are roughly 10 meters away from each other, covering in one of the target sites, Radvanice, an area of 70 x 70 meters. Over the years the trees will be growing over the sensor boxes, enabling the assessment of the filtration efficiency of canopies. Continuous measurements are performed by the sensor network at the target sites of CLAIRO, the values being recorded every 10 seconds. This results in an enormous amount of data. In 20 months of measurements the database of CLAIRO has grown to have 28 million rows. See additional details in CLAIRO Zoom-in 1.

The sensor units had been installed before the greenery was planted so that the consortium could compare the impacts of the original and the newly planted vegetation on air quality. The measurements have started in September 2019, and data was already gathered for more than a year until spring 2021 when the planting of the new greenery took place. This way baseline data could be collected for all seasons. The measurements are planned for 8 years altogether. They will continue well after the closing of the project, so that with the growing vegetation also the changes in the local concentration of air pollutants can be followed, and pollution removal by trees and shrubs can be assessed.

Apart from making possible the estimation of pollutant capture by green infrastructure, in the initial phase of the project air quality measurements also supported the design of the most effective composition and structure of urban green spaces, enabling the optimization of tree and shrub planting. Through the monitoring activity it is possible to determine the vegetation area that can effectively mitigate exposure to air pollutants.

In the sensor units, concentrations of particulate matter, gaseous pollutants (nitrogen dioxide and ozone) and volatile organic compounds (VOC) are monitored by separate sensors. 

The data collected by the sensors are processed and stored by the Floreon geo-database developed by VSB. The database that can be accessed at the airsens.eu website allows data collection, the calibration of the sensors, data visualization (the selection of pollutants, sites and time periods) and data evaluation.

Long-term measurements provide an exceptional opportunity

The results of the CLAIRO project are quite promising, as air quality measurements have shown that the planted greenery seem to be effective in capturing air pollutants. Pavel Bucek, from the Technical University of Ostrava, added details:

We have compared the long-term concentration profiles to the predicted models of the capture, and so far, because the greenery is young and growing, we are at approximately 30% of capture and the trendlines are following the model.

See additional details about the model of capture of air pollutants developed under the project in CLAIRO Zoom-in 3. 

Although the sensor technology is developing rapidly, there is still a general lack of information on measurement validity of sensors and on their behaviour in case of long-term measurements, exceeding one year.

CLAIRO presented a large set of valid results that can help understanding the limitations of sensory measurements, and in addition data was available for a period longer than 18 months. The researchers at the Technical University of Ostrava decided to use this huge amount of data to their advantage and grabbed the opportunity to assess the overall applicability of the sensor technology and to check the useful lifetime of the sensors.

Pavel Bucek, referring to the monitoring activity performed under CLAIRO, stated that “We process a lot of data, so we try to use some data mining technics on that to see if it can help us in the long-term.” The researchers at VSB wanted to find out how and when sensor data can be used without sophisticated analytical techniques.

Talking about the results of the analysis, Pavel Bucek has highlighted:  

We are frankly surprised that the sensors have a service life longer than expected. We expected that we would need to change some of the sensors in the boxes and so far, all of them are still working and it seems that they produce numbers similarly as at the beginning of the project.

The sensor system was found to be quite robust and stable. Sensory measurements of particulate matter were fairly reliable, robustness decreased only at low concentration levels. Data quality in case of NO2 measurements were lower, however concentration changes were in line with reference measurements and raw data could be improved by careful calibration of the sensors and more advanced data analysis technics. In case of VOC measurements, sensors exhibited a lack of selectivity, as they did not support the identification of individual pollutants.

In general, fast responses to concentration changes, a high resolution in space and time, unnecessary sampling and laboratory analysis, as well as low costs give sensor networks an edge over conventional monitoring instruments. However sensors cannot yet be used for regulatory or compliance purposes. Still, as findings of VSB indicate, they can be a source of high-density complementary data that can help monitoring trends and enable prediction of changes in air quality. As evidences indicate, sensor networks can be efficient tools for indicative air quality measurements.

As Pavel Bucek has explained many air quality professionals are still complaining about the quality of data provided by sensors. Therefore, researchers at VSB try to get across the message that although the quality of data from a sensor that costs only 20 USD will never be the same as the one provided by professional technology, but the low price to purchase a single professional monitoring station, a lot of sensors can be bought. On top of this, the benefits of a network can also be enjoyed.

References

Bilek, J., Bilek, O., Marsolek, P., Bucek, P. (2021) Ambient Air Quality Measurement with Low-Cost Optical and Electrochemical Sensors: An Evaluation of Continuous Year-Long Operation. Environments. 8. 114. 10.3390/environments8110114.

Bilek, J., Marsolek, P., Bilek, O., Bucek, P. (2022) Field Test of Mini Photoionization Detector-Based Sensors—Monitoring of Volatile Organic Pollutants in Ambient Air. Environments. 9. 49. 10.3390/environments9040049.

Bucek, P., Marsolek, P., Bilek, J. (2021). Low-Cost Sensors for Air Quality Monitoring - the Current State of the Technology and a Use Overview. Chemistry-Didactics-Ecology-Metrology. 26. 41-54. 10.2478/cdem-2021-0003.

Zapletal, M., Vit, K., et al. (2021) Green Infrastructure and its Effect on Air Quality: Methodology of planting greenery in urban areas in order to capture pollution. Statutory City of Ostrava, Strategic Development Department. 978-80-88399-04-9.

https://aaqr.org/articles/aaqr-21-03-oa-0073

http://www.cas.manchester.ac.uk/restools/instruments/aerosol/opc/